Feasibility of continuous glucose monitoring in critically ill emergency department patients

Background: Glucose control is important in the management of critically ill patients. However, strict glucose control requires a large amount of nursing resources, especially in overcrowded emergency departments (EDs).

Objectives: A continuous glucose monitoring system (CGMS) may be beneficial for glucose control in the ED. The objective of this study was to determine the test characteristics of CGMS in critically ill ED patients.

Methods: A prospective observational study of critically ill ED patients was conducted. During a patient's visit to the ED, a CGMS sensor measured their interstitial fluid glucose levels continuously. Capillary glucose was measured every hour and used for glucose control and as a reference value. CGMS values were recorded in real time and compared with capillary glucose values.

Results: A total of 122 pairs of capillary and CGMS glucose values in 12 patients were analyzed. The correlation coefficient was 0.87, and Bland-Altman analysis showed that 117 pairs (95.9%) were within a 95% confidence interval. A Clarke Error Grid Analysis indicated an overall accuracy of 96.8% (Zones A and B). However, the mean absolute relative difference (MARD) was significantly higher in the hypoglycemic range than in a normo- or hyperglycemic range (p = 0.001). The sensitivity and positive predictive value of CGMS for detecting hypoglycemia were 33.3% and 16.7%, respectively. The CGMS specificity and negative predictive value were 95.8% and 98.3%, respectively. There was no linear correlation between MARD and body mass index, axillary temperature, inotrope score, and base deficit (all p-value >0.05).

Conclusion: CGMS demonstrated good clinical accuracy by Clarke Error Grid Analysis. There also was high agreement between CGMS and capillary glucose levels. However, CGMS demonstrated only limited real-time hypoglycemia detection ability in critically ill ED patients.